We propose to continue our research group?s work to identify and validate hormonal markers that predict risk of invasive breast cancer in postmenopausal women. Although the estrogen pathway plays a critical role in breast carcinogenesis, accruing evidence suggests measurement of the classic estrogens only, as done in prior epidemiologic studies (including our own) may miss a sizeable component of estrogenic activity that is etiologically relevant to risk. Further, the insulin pathway is another critical, but relatively understudied, hormonal pathway that may impact breast cancer risk. Using a prospective nested case-control design, we plan to analyze blood samples collected from the 32,826 participants in the Nurses' Health Study (NHS) who provided a blood sample in 1989-90 and, for 18,743 of these women, a second sample in 1999-2000. Specifically, we propose to evaluate the independent associations of 27-hydroxycholesterol (27HC; a recently identified endogenous selective estrogen receptor modulator) as well as an estrogen bioassay (that assesses estrogen pathway activation) in relation to breast cancer risk. We also will evaluate several new aspects of the relationship between plasma c-peptide (reflecting insulin secretion) and breast cancer risk. With 26 years of follow-up and two blood samples in a subset of women, we will evaluate the importance of timing of c-peptide exposure in breast carcinogenesis. Further, we will assess whether the c-peptide/breast cancer relationship varies by breast tumor expression of insulin receptor or several markers of PI3k/Akt/mTOR pathway activation. Previous work, by not accounting for these tumor molecular characteristics, may have underestimated the etiologic importance of insulin in breast cancer risk and hence opportunities for prevention. We also will evaluate if c-peptide (and secondarily the novel estrogen related markers) improve the Gail and Rosner-Colditz risk prediction models, and validate the addition of multiple biomarkers (i.e., the plasma hormones, a genetic risk score and mammographic density) to these models in the prospective Mayo Mammography Health Study (MMHS) and Melbourne Collaborative Cohort Study (MCCS). The MMHS and MCCS have risk factor and biomarker information that is comparable to that in the NHS. In addition to providing one or two blood samples, NHS cohort members have provided detailed exposure and disease information biennially (including data on most known or probable breast cancer risk factors) since 1976, and, for about 70% of breast cancer cases, we have collected formalin-fixed paraffin embedded tumor tissue, thus allowing molecular analyses of the tumor. The recently funded NHS UM1 grant will provide cohort follow-up, breast cancer documentation, and important covariate data, increasing the cost-effectiveness of the proposed project. This project, by working across the translational research spectrum from discovery of novel hormonal biomarkers to practical application of well confirmed biomarkers in risk prediction models, should add considerable insight into breast cancer etiology and ways to better identify high risk women who may benefit from risk reduction efforts (e.g., chemoprevention).